Setting Up NTP on a Server: A Step-by-Step Guide
Servers
19.11.2024
Reading time: 9 min
Share
NTP (Network Time Protocol) is used to synchronize system time with a reference time provided by special servers. This article will cover how to configure NTP on various operating systems and devices, starting with a comprehensive guide on setting up an NTP Server on Linux.
Configuring an NTP Server on Linux
We'll demonstrate synchronization setup using Ubuntu, but this guide also applies to Debian and most Linux-based systems. We’ve divided the instructions into three parts: the first covers installing the NTP server, the second explains synchronizing NTP clients, and the third covers advanced synchronization settings.
To follow this guide, you will need:
- A cloud server with Ubuntu installed
- A
rootuser or a user withsudoprivileges nanoor any other editor installed
Installing the NTP Server
These steps will guide you through installing and preparing the NTP server for further configuration.
- Update the repository index to ensure you can download the latest software versions. Use the following command:
sudo apt-get update- Install the NTP server:
sudo apt-get install ntpConfirm the installation by choosing Y if prompted (Y/N). Wait until the software is downloaded and installed.
- Verify the installation:
sntp --versionThe output should display the version number and the installation time.
- Switch to the nearest server pool.
The server should receive accurate time by default, but it’s better to connect to a server pool closest to your location for extra reliability. To do this, edit the ntp.conf file located at /etc/ntp.conf. Open it with nano (you need sudo privileges) by entering:
sudo nano /etc/ntp.confYou’ll see four lines, which we’ve highlighted in orange for reference:
These are the default pools, which we’ll replace with local ones (for example, for the USA, we can use NTP servers from this page). After replacing the lines, save and close ntp.conf by pressing Ctrl+O and Ctrl+X.
- Restart the server:
sudo service ntp restart- Check the server status:
sudo service ntp statusThe output should indicate Active (running) on one of the first lines (Active) and the server start time.
- Configure the firewall.
To allow client access to the server, open UDP port 123 using UFW with the following command:
sudo ufw allow from any to any port 123 proto udpThe installation is complete, and the server is running; now, you can proceed with further configuration.
Configuring NTP Client Synchronization
The following steps will allow client systems to synchronize with our NTP server, which will serve as their primary time source.
- Check the Connection
To verify the network configuration for NTP, enter the following command in the terminal:
sudo apt-get install ntpdate- Specify IP Address and Hostname
To configure the server’s IP and hostname, edit the hosts file located at /etc/hosts:
sudo nano /etc/hostsAdd the relevant data in the third line from the top (the address below is just an example; replace it with the actual IP of your NTP server):
192.168.154.142 ntp-serverPress Ctrl+X to exit and save changes by pressing Y. Alternatively, if you have a DNS server, you can perform this step there.
- Verify Client Synchronization with the Server
To check if synchronization is active between the server and client, enter:
sudo ntpdate ntp-serverThe output will show the time offset. A few milliseconds difference is normal, so you can ignore small values.
- Disable the timesyncd Service
This service synchronizes the local system time, but we don't need it since our clients will sync with the NTP server. Disable it with:
sudo timedatectl set-ntp off- Install NTP on the Client System
Install NTP on the client with this command:
sudo apt-get install ntp- Set Your NTP Server as the Primary Reference
To ensure clients sync specifically with your server, open the ntp.conf file and add the following line:
server NTP-server-host prefer iburstThe prefer directive marks the server as preferred, and iburst allows multiple requests to the server for higher synchronization accuracy. Save the changes by pressing Ctrl+X and confirming with Y.
- Restart the Server
Restart the NTP server with this straightforward command:
sudo service ntp restart- Check the Synchronization Queue
Finally, check the synchronization status by entering:
ntpq -psThis command displays the list of servers in the synchronization queue, including your NTP server as the designated source.
Advanced Synchronization Options
Now that we’ve set up the NTP server and synchronized client machines, we’ll revisit the ntp.conf file (located at /etc/ntp.conf), which contains additional configurations to ensure robust synchronization with external sources.
Preferred Server
Mark the most reliable server or server pool with the prefer directive we’ve used before. For example:
server 1.north-america.pool.ntp.org preferThe server directive indicates a specific server, while pool can be used to specify a pool of servers. Don’t forget the line server 127.127.1.0 at the end of the pool list, which defaults to the system time if the connection is lost.
Security Settings
Make sure the following lines are included in ntp.conf:
restrict default kod notrap nomodify nopeer noqueryThe default command applies these settings as defaults for all restrict commands:
-
kod(Kiss-o’-Death) limits the rate of requests. -
notrapblocks the acceptance of control commands. -
nomodifyrestricts commands that might alter the server state. -
nopeerprohibits synchronization with external hosts. -
noqueryblocks query requests.
For IPv4, use -4 before default, and for IPv6, use -6.
Here’s an example of using some of these commands. The following line allows synchronization of nodes in a specific network while restricting nodes from receiving control or state-altering commands:
restrict 192.168.0.0 mask 255.255.255.0 nomodify notrapThe following lines are required for the server to communicate with itself:
restrict 127.0.0.1
restrict ::1Finally, remember to restart the server after making these changes.
Verifying NTP Operation
To check if NTP is functioning correctly, use the command ntpq -p. Example output:
In the first column, you’ll see the synchronization server’s address, followed by its parent server, stratum level (st column), and nup (t column). The next three columns show details on the last synchronization time, sync interval, and reliability status . The final two columns display the time difference between the synchronized server and the reference server, as well as the offset.
Pay attention to the symbols in the first column, which appear before the IP address:
- A
+symbol indicates a reliable server for synchronization and a-means the opposite. - An
*indicates the current server chosen for synchronization. - Occasionally, an
xwill appear, which means the server is unavailable.
Checking if the Server Provides Accurate Time
To ensure the server is distributing the correct time, run the ntpdate command from another system, specifying the IP address of the NTP server you want to verify. The output should look something like this:
adjust time server (IP address here) offset 0.012319 secThe number represents the time offset. Here, an offset of about 0.01 seconds (12 milliseconds) is perfectly acceptable.
Now that we’ve completed the Linux setup, let’s look at configuring the NTP protocol on Windows.
Configuring an NTP Server on Windows Server
To install and configure an NTP server on Windows Server, you'll need to make some changes in the registry and run commands in the command prompt.
Before proceeding with the configuration, you must start the service. This is done by modifying the following registry entry:
HKLM\System\CurrentControlSet\services\W32Time\TimeProviders\NtpServerIn this section, find the Enabled entry on the right and set it to 1 so that the Data column displays:
0x00000001 (1)Next, open cmd and enter the command needed to restart the protocol:
net stop w32time && net start w32timeMake sure to run this command from C:\Users\Administrator. To verify that NTP is enabled, use the following command:
w32tm /query /configurationYou’ll get a long entry, and you should check the block NtpServer <Local>. In the Enabled line, the value should be 1. Now, open UDP port 123 in the firewall for proper client servicing, and then proceed with the configuration.
Return to the registry and look for the entry:
HKLM\System\CurrentControlSet\services\W32Time\ParametersThis section contains many parameters, but the main one is Type, which can take one of four values:
NoSync— No synchronization.NTP— Synchronization with external servers specified in the NtpServer registry entry (this is the default for standalone machines).NT5DS— Synchronization according to the domain hierarchy (default for machines in a domain).AllSync— Synchronization with all available servers.
Now, go back to the registry and configure the values under the NtpServer section. Most likely, only the Microsoft server is listed. You can add others, paying attention to the flag at the end:
- 0x1, SpecialInterval — Standard mode recommended by Microsoft.
- 0x2, UseAsFallbackOnly — Use this server as a fallback.
- 0x4, SymmetricActive — This is the main mode for NTP servers.
- 0x8, Client — Used when synchronization issues occur.
The last thing you need to do is set the synchronization interval in the section:
W32Time\TimeProviders\NtpClientThe parameter is SpecialPollInterval, where you should set the desired value (in seconds). By default, it’s set to one week. If you want more frequent synchronization, set:
- 86400 for 1 day.
- 21600 for a quarter of a day (6 hours).
- 3600 for 1 hour.
The last value is optimal in terms of system load and acceptable precision when frequent synchronization is required.
Configuring an NTP Server on Cisco Devices
On Cisco devices, the process is simple and quick:
-
Enter configuration mode with the command:
conf t-
Set the time zone using the command:
clock timezone <timezone> <offset>For example:
clock timezone CST -6-
Next, enter the command to set the NTP source:
ntp source-
Specify the source.
-
If you want to make the server the primary one for other machines in the network, use the following command:
ntp master 2The number should be 2 or greater.
-
Use the command
ntp update-calendarto update the time. -
Enter the names or IP addresses of the NTP servers.
-
Enter the time zone with the command:
clock timezone-
And set the source using:
ntp sourceTo check the configuration or troubleshoot, use the show command. It will be useful for checking the time (show clock), NTP status (show ntp status), and associations (show ntp associations).
Configuring an NTP Server on MikroTik Routers
We will configure the NTP server using SNTP:
-
In Winbox, go to System – SNTP Client.
-
Find the SNTP Client section and enable it by checking the Enabled box.
-
In the Server DNS Names field below, enter the IP addresses of the NTP servers.
-
To check if everything is working, go to System – Clock. Set the time zone by choosing it from the dropdown list or check the Time Zone Autodetect box, and the time zone will be set automatically.
-
The synchronization interval can be seen in the Poll Interval field in the SNTP Client menu. Below, you will find the last synchronization time in the Last Update field.
That’s it! Now you’ve learned how to configure NTP on different operating systems and devices.
Servers
19.11.2024
Reading time: 9 min
Share
Similar
Servers
How to Protect a Server from DDoS Attacks
A DDoS attack (Distributed Denial of Service) aims to overwhelm a network with excessive traffic, reducing its performance or causing a complete outage. This is reflected in the term "denial-of-service" (refusal of service).
The frequency and intensity of DDoS attacks have been rising rapidly. A report by Cloudflare noted that in 2021, the number of attacks grew by one-third compared to 2020, with a peak in activity observed in December.
The duration of a DDoS attack can vary. According to research by Securelist:
94.95% of attacks end within four hours.
3.27% last between 5 to 9 hours.
1.05% persist for 10 to 19 hours.
Only 0.73% of all attacks extend beyond 20 hours.
Effective Tools for Protecting a Server from DDoS Attacks
If you don't want to rely on vendors' solutions, paid services, or proprietary software, you can use the following tools to defend against DDoS attacks:
IPTables. A powerful firewall tool available in Linux systems that allows precise control over incoming and outgoing traffic.
CSF (ConfigServer Security and Firewall). A robust security tool that simplifies managing firewall rules and provides additional protection mechanisms.
Nginx Modules. Modules specifically designed for mitigating DDoS attacks, such as limiting the number of requests per IP or delaying excessive requests.
Software Filters. Tools or scripts that analyze and filter traffic to block malicious or excessive requests, helping to maintain service availability.
IPTables. Blocking Bots by IP Address
The IPTables tool helps protect a server from basic DDoS attacks. Its primary function is to filter incoming traffic through special tables. The resource owner can add custom tables.
Each table contains a set of rules that govern the tool's behavior in specific situations. By default, there are only two response options: ACCEPT (allow access) and REJECT (block access).
In IPTables, it is possible to limit the number of connections. If a single IP address exceeds the allowed number of connections, the tool will block access for that IP. You can extend the tool's functionality with additional criteria:
Limit: Sets a limit on the number of packet connections within a chosen time period.
Hashlimit: Works similarly to Limit, but applies to groups of hosts, subnets, and ports.
Mark: Used to mark packets, limit traffic, and filter.
Connlimit: Limits the number of simultaneous connections for a single IP address or subnet.
IPRange: Defines a range of IP addresses that are not considered as a subnet by the tool.
Additionally, IPTables can use criteria such as Owner, State, TOS, TTL, and Unclean Match to set personalized configurations, effectively protecting the resource from DDoS attacks.
The ipset kernel module allows you to create a list of addresses that exceed the specified connection limit. The ipset timeout parameter sets a time limit for the created list, which is enough to ride out a DDoS attack.
By default, IPTables settings return to their basic configuration after a system reboot. To save the settings, you can use additional utilities (such as iptables-save or iptables-persistent), but it is recommended to start with the default options to avoid saving incorrect settings that could block server access for everyone.
ConfigServer Security and Firewall
While IPTables is a convenient and effective tool, it can be quite complex to configure. You’ll need to learn how to manage it and write additional scripts, and if something goes wrong, your resource may end up being a "closed club" for just a few users.
CSF (ConfigServer Security and Firewall) is a "turnkey" configurator, meaning you only need to set the correct parameters and not worry about the server's security.
Installing the Server Firewall
The preliminary installation steps involve downloading two additional components required to run CSF: the Perl interpreter and the libwww library. The next step is to install ConfigServer Security and Firewall itself. Since the tool is not available in the official repository, you'll need to download it directly from the provided link or by fetching the ready-made archive:
cd /usr/srcwget https://download.configserver.com/csf.tgz
After downloading, extract the archive and move it to the defender’s files folder. Then, run the installation process. Once installed successfully, you can proceed with configuring CSF.
Configuring the Server Firewall
By default, the settings in ConfigServer and Firewall are active for 5 minutes, after which any changes are reset. This test format is useful for conducting experiments and understanding errors in the applied configuration. To switch to live mode, change the Testing value to 0.
Proper configuration of CSF ensures reliable protection against DDoS attacks. Here are some essential commands in CSF:
Specify incoming ports:
TCP_IN = "22,23,25,36,75,87"
Specify outgoing ports:
TCP_OUT = "22,23,25,36,75,87"
Configure email notifications for SSH connections:
LF_SSH_EMAIL_ALERT = "1"
Add an IP address to the exception list (useful for server management teams):
csf -a 192.168.0.7
Block a specific IP address from connecting to the server:
csf -d 192.168.0.6
Nginx Modules
How can you protect your server from DDoS attacks using simpler methods? Use Nginx modules like limit_conn and limit_req. The limit_conn module limits the maximum number of connections to the server, while the limit_req module limits the number of requests within a specified time frame.
For example, if you want to limit the number of simultaneous connections to 30 and restrict the number of connections within a 3-second window, the configuration will look as follows:
limit_conn_zone $binary_remote_addr zone=perip: 30m;limit_req_zone $binary_remote_addr zone=dynamic:30m rate=3r/s;
This configuration allows only 3 requests per second. Any additional requests are queued. The burst parameter controls the queue size. For example, if the burst value is set to 7, the module will queue up to 7 requests when the request count exceeds 10, while any further requests will be rejected with an error.
Software Filter
Server protection against DDoS attacks can also be achieved using web applications. A traffic filter uses JavaScript, which is inaccessible to bots, effectively redirecting DDoS attacks to a placeholder page.
The operation of the filter is simple. The configuration defines conditions for blocking bots, and when a visitor meets those conditions, they are redirected to a placeholder page instead of the requested page. The filter can also specify the reason for the redirection.
03 December 2024 · 6 min to read
Servers
How to Protect a Server: 6 Practical Methods
Any IT infrastructure requires robust protection. While information security is a vast topic, there are basic steps that can safeguard against attacks from amateur hackers and bots. This article outlines six straightforward methods to protect your server effectively.
Tools and Methods of Protection
Securing a server from breaches involves a combination of measures. These can be categorized into the following areas:
Securing communication channels used for system administration and operation.
Implementing multi-layered security for the system.
Restricting access to infrastructure resources.
Monitoring and auditing system activities.
Backing up data.
Timely updates or rollbacks of software.
Antivirus protection for servers.
Below, we detail six practical methods to achieve a robust security level against amateur attackers and bots.
Privilege Restriction
When managing access to resources, follow the principle of least privilege: users and processes should only have the minimal permissions necessary to perform their tasks. This is particularly important for databases and operating systems. This approach not only prevents unauthorized external access but also mitigates risks from internal threats.
Separate Accounts for Administrators: Create individual accounts for each admin. Use non-privileged accounts for operations that don’t require elevated permissions.
Active Directory: In environments using Microsoft Active Directory, regularly audit and configure group policies. Mismanagement of these policies can lead to severe security breaches, especially if exploited by a malicious admin or hacker.
Minimize Root Usage in Unix Systems: Avoid working as the root user. Instead, disable the root account and use the sudo program for tasks requiring elevated permissions.
To customize sudo behavior, modify the /etc/sudoers file using the visudo command. Below are two useful directives for monitoring sudo activity.
By default, sudo logs to syslog. To store logs in a separate file for better clarity, add the following to /etc/sudoers:
Defaults log_host, log_year, logfile="/var/log/sudo.log"
This directive records command logs, along with input and output (stdin, stdout, stderr), into /var/log/sudo-io:
Defaults log_host, log_year, logfile="/var/log/sudo.log"
For a deeper dive into managing the sudoers file, check this guide.
Mandatory Access Control (MAC)
This recommendation focuses on Linux systems and builds upon the principle of access control. Many Linux administrators rely solely on discretionary access control (DAC) mechanisms, which are basic and always active by default. However, several Linux distributions include mandatory access control (MAC) mechanisms, such as AppArmor in Ubuntu and SELinux in RHEL-based systems.
While MAC requires more complex configuration of the OS and services, it allows for granular access control to filesystem objects, significantly enhancing the server's security.
Remote Administration of Operating Systems
When remotely administering an operating system, always use secure protocols:
For Windows, use RDP (Remote Desktop Protocol).
For Linux, use SSH (Secure Shell).
Although these protocols are robust, additional measures can further strengthen security.
For RDP, you can block connections of accounts with blank passwords. You can configure it via Local Security Policy under the setting:
Accounts: Limit local account use of blank passwords to console logon only.
RDP sessions can be protected with the secure TLS transport protocol, which will be discussed later.
By default, SSH user authentication relies on passwords. Switching to SSH key-based authentication provides stronger protection, as a long key is far more difficult to brute-force than a password. Additionally, key-based authentication eliminates the need to enter a password during login since the key is stored on the server.
Setting up keys requires only a few simple steps:
Generate a key pair on your local machine:
ssh-keygen -t rsa
Copy the public key to the remote server:
ssh-copy-id username@remote_address
If key-based authentication is not an option, consider implementing Fail2ban. This tool monitors failed login attempts and blocks the IP addresses of attackers after a specified number of failed attempts.
Additionally, changing default ports can help reduce the likelihood of automated attacks:
Default SSH port 22/tcp → Choose a non-standard port.
Default RDP port 3389/tcp → Use a custom port.
Firewall Configuration
A robust security system is layered. Relying solely on access control mechanisms is insufficient; it is more logical to manage network connections before they reach your services. This is where firewalls come in.
A firewall provides network-level access control to segments of the infrastructure. The firewall decides which traffic to permit through the perimeter based on a specific set of allow rules. Any traffic that does not match these rules is blocked. In Linux, the firewall is integrated into the kernel (via netfilter), and you can manage using a frontend tool such as nftables, iptables, ufw, or firewalld.
The first step in configuring a firewall is to close unused ports and keep only those that are intended for external access. For instance, a web server typically requires ports 80 (HTTP) and 443 (HTTPS) to remain open. While an open port itself is not inherently dangerous (the risk lies in the program behind the port), it is still better to eliminate unnecessary exposure.
In addition to securing the external perimeter, firewalls can segment infrastructure and control traffic between these segments. If you have public-facing services, consider isolating them from internal resources by using a DMZ (Demilitarized Zone). Additionally, it’s worth exploring Intrusion Detection and Prevention Systems (IDS/IPS). These solutions work on the opposite principle: they block security threats while allowing all other traffic through.
Hostman offers a cloud firewall that provides cutting-edge defense for your server.
Virtual Private Networks (VPNs)
Up until now, we have focused on protecting a single server. Let’s now consider securing multiple servers. The primary purpose of a Virtual Private Network (VPN) is to provide secure connectivity between organizational branches. Essentially, a VPN creates a logical network over an existing network (e.g., the Internet). Its security is ensured through cryptographic methods, so the protection of connections does not depend on the underlying network's security.
There are many protocols available for VPNs, and the choice depends on the size of the organization, network architecture, and required security level.
PPTP (Point-to-Point Tunneling Protocol) is a simple option for a small business or home network, as it is widely supported on routers and mobile devices. However, its encryption methods are outdated.
For high-security needs and site-to-site connections, protocols like IPsec are suitable.
For site-to-host connections, options like WireGuard are more appropriate. WireGuard and similar protocols provide advanced security but require more intricate configuration compared to PPTP.
TLS and Public Key Infrastructure (PKI)
Many application-layer protocols, such as HTTP, FTP, and SMTP, were developed in an era when networks were limited to academic institutions and military organizations long before the invention of the web. These protocols transmit data in plaintext. To ensure the security of a website, web control panels, internal services, or email, you should use TLS.
TLS (Transport Layer Security) is a protocol designed to secure data transmission over an untrusted network. While the term SSL (e.g., SSL certificates, OpenSSL package) is often mentioned alongside TLS, it’s important to note that the modern versions of the protocol are TLS 1.2 and TLS 1.3. Earlier versions of TLS and its predecessor, SSL, are now considered obsolete.
TLS provides privacy, data integrity, and resource authentication. Authentication is achieved through digital signatures and the Public Key Infrastructure (PKI). PKI functions as follows: the server's authenticity is verified using an SSL certificate, which is signed by a Certificate Authority (CA). The CA’s certificate is, in turn, signed by a higher-level CA, continuing up the chain. The root CA certificates are self-signed, meaning their trust is implicitly assumed.
TLS can also be used with Virtual Private Networks (VPNs), such as setting up client authentication using SSL certificates or a TLS handshake. In this case, it would be necessary to organize your own PKI within the local network, including a CA server, as well as the keys and certificates for network nodes.
The Dangers of Attackers
The level of threat depends on the type of attack. Cyberattacks can be broadly categorized into two main types.
Breaching the Security Perimeter
This type of attack involves gaining unauthorized access to the account of an authenticated user of a service or system, such as a database. Breaches of privileged accounts pose significant risks because attackers gain the ability to view sensitive information and modify system parameters. The most critical type of breach involves gaining unauthorized access to the superuser account of the operating system, potentially compromising a significant portion of the infrastructure.
Disabling Systems
This category of attacks aims to disrupt system operations rather than steal data, but it is no less dangerous. The most prominent example is a DoS (Denial of Service) or DDoS (Distributed Denial of Service) attack. These attacks overload the server with a flood of requests, causing it to fail and become unresponsive to legitimate users. In some cases, a DoS attack serves as a precursor to other forms of cyberattacks.
The results of cyberattacks often include data breaches, financial losses, and reputational damage. For this reason, even the most basic level of security should be implemented when establishing an IT infrastructure.
02 December 2024 · 8 min to read
Servers
Load Testing: Objectives, Tasks, and Procedure
This article explores the features and benefits of load testing a web server, discussing why it is important and how to perform it correctly.
What Is Load Testing?
Load testing is the process of evaluating the performance and reliability of a web server using specialized tools designed to simulate real-world server loads. These tools emulate the activity of a specified number of users and document the resulting load on the server. The collected data is then analyzed to assess the performance of hardware resources, communication channels, and server software.
Why Use Load Tests: Objectives of Testing
Most websites and applications are created to generate revenue, or profitability is set as one of the project goals. The performance of the server—its ability to handle the planned number of simultaneous users—is a key success factor.
If a server cannot handle a surge in visitors, it results in decreased traffic, negatively impacting the website's behavioral metrics. As a result, the site's ranking in search engine results drops, reducing organic traffic and leading to a decline in sales and advertising revenue. Such failures can be equally disastrous for web applications used by thousands of people.
The primary goal of load testing is to evaluate server capacity under extreme conditions, pushing it to its operational limits. This helps determine whether additional resources are needed or if existing ones are sufficient for stable operation. The outcome includes mitigating the risk of site or application downtime and achieving significant cost savings in the long run.
Step-by-Step Guide to Load Testing a Server
Let’s break down the entire process into sequential steps:
Preparation for Testing. Before conducting load testing, start with functional testing to ensure the chosen tools and configurations are correct.
Define Objectives. Typical objectives include identifying the server’s performance limits and detecting system bottlenecks.
Specify Requirements. Clearly define the requirements, such as: 90% of users must be served within a maximum of 10 seconds each.
Develop Scenarios. Create scenarios based on typical user behavior on the website, application, or service.
Choose Tools. Select software that best aligns with the testing goals.
Configure Tools. Set the load levels and write scripts to simulate user behavior.
Execute Testing. Gradually increase the load while documenting critical thresholds.
Analyze Results. Process the collected data, draw conclusions, and prepare recommendations for improving system performance.
Objectives and Requirements
The type and scale of the load, as well as the metrics to monitor, depend on the specific objectives. Common tasks include:
Determining the server’s performance limits.
Checking configuration reliability.
Monitoring backups.
Identifying problematic areas in the system.
Regarding requirements, they often define user service times as percentages. It’s important to avoid aiming for 100% of users to be served within a strict timeframe, as a buffer (typically around 10%) is necessary. This allows the system to handle unexpected events without failures.
User Scenarios
User scenarios depend on how users interact with the site. For example, a typical scenario for an online store might include:
Logging in.
Searching for and selecting a product.
Viewing the product details.
Adding the product to the cart.
Proceeding to the cart.
Initiating the checkout process.
Filling in form fields.
Confirming and paying for the purchase.
The exact flow depends on the functionality of the site or application. After modeling one or more typical scenarios, identify the most resource-intensive pages and select tools to simulate the load on these critical points.
Tools for Load Testing
If the objectives allow, it is reasonable to use free and open-source tools for testing. One of the most popular options is Apache JMeter, a highly configurable cross-platform software that supports all web protocols. JMeter makes it easy to develop scripts that simulate user actions on a website or application. Once the scripts are created, we can set the load levels and proceed with the testing process.
However, JMeter is not the only tool for load testing. Other options include WAPT, NeoLoad, Siege, Gobench, WRK, Curl-loader, Tsung, and many more. Each of these tools has unique features. Before choosing one, review their descriptions, study available information, and consider user reviews and forums.
Load Testing
After defining typical scenarios and selecting appropriate tools, the testing process begins. Most scenarios involve gradually increasing the load. The number of concurrent threads or users increases until response times rise. This marks the first critical threshold, often referred to as the degradation point.
The second threshold, known as the sub-critical point, occurs when response times exceed acceptable limits. The system can still process requests at this stage, but response times hit the SLA (Service Level Agreement) threshold. Beyond this point, delays accumulate rapidly, causing the system to reach the critical point.
The critical point, or failure point, occurs when the server's resources are exhausted—either CPU power or memory runs out. At this stage, the server crashes, signaling the end of testing and the start of data analysis.
Analysis of Load Testing Results
Testers analyze the collected data to identify bottlenecks. Sometimes, you can resolve the issues by adjusting configurations or refining the code. In other cases, a specific service within the project may cause delays, requiring targeted optimization. This might involve configuration adjustments or scaling the service.
For high user volumes, the most common issue is hardware overload. Typically, addressing this requires upgrading the infrastructure—for example, adding RAM or switching to a more powerful processor.
Conclusion
Load testing a server is an essential procedure for anyone looking to avoid failures in a growing website, service, or application. Practical experience shows that proper configuration adjustments or code optimization can significantly enhance server performance. However, to achieve these improvements, it’s critical to identify system bottlenecks, which is precisely the purpose of load testing.
02 December 2024 · 5 min to read
Do you have questions,
comments, or concerns?
Our professionals are available to assist you at any moment,
whether you need help or are just unsure of where to start.
whether you need help or are just unsure of where to start.
